Cover Page Improving Delivery of Patient-Centered Cardiac Rehabilitation NCT02105246 April 12, 2018

Study Protocol Background Ischemic heart disease (IHD) is the leading cause of death in Veterans and has been identified as a priority condition for improving healthcare within the VA. Exercise-based cardiac rehabilitation (CR) is an evidence-based, cost effective therapy that reduces morbidity and mortality following acute myocardial infarction (MI), coronary artery bypass grafting (CABG) or percutaneous coronary intervention (PCI). The American Heart Association (AHA) and American College of Cardiology (ACC) recommend referral to CR as one of its 9 performance measures for secondary prevention (along with aspirin, beta blockers, statins and smoking cessation) in patients with IHD. Unfortunately, despite the compelling benefits and widespread endorsement of its use, CR is vastly underutilized in VHA. A recent Presidential Advisory from theAHA concluded, “The remarkably wide treatment gap between scientific evidence of the benefits of cardiac rehabilitation and clinical implementation is unacceptable.” Only 35 (27%) of VA facilities offer CR programs, and less than 10% of eligible veterans receive this guideline-recommended therapy. Geographic distance is by far the largest barrier to participation. Of the 9.2 million veterans currently enrolled in VHA, 6.7 million (73%) live more than 60 minutes from a VA CR center [based on geography data from VA Planning Systems Support Group (PSSG)]. Thus, there is an urgent need to adapt CR programs to improve access to and utilization of CR among rural veterans. In a recent study conducted at the Iowa City VA (Wakefield et al, 2014), 48 patients completed a home-based program and 12 patients completed face-to-face CR. There was no difference in blood pressure, lipid levels, weight, glycosylated hemoglobin or rate of rehospitalization between the two groups. Home-based CR participants were highly satisfied with their care and had a better completion rate than patients who underwent face-to-face CR (89% vs. 73%). Costs for home and center-based CR programs were comparable. These findings suggest that a home-based CR program is a viable, safe, and cost- effective alternative to center-based programs. In a 2010 Cochrane review, there was no significant difference in mortality among patients randomly assigned to home-based CR vs. center-based CR. Objectives   Test the hypothesis that patients undergoing home based cardiac rehabilitation will not differ in 6 minute walk time, physical activity, quality of life, patient and caregiver satisfaction, and re-hospitalization as compared with patients undergoing center-based cardiac rehabilitation.  Aim 1. Determine whether automatic referral to home- vs. center-based CR increases patient participation in CR after hospitalization for myocardial infarction or coronary revascularization. Aim 2. Among patients who choose to participate in CR, compare the effectiveness of home- vs. center-based CR on six-minute walk distance, quality of life, and healthcare expenditures. Aim 3: Determine whether the effects of home vs. center-based CR differ by age, gender, race, ethnicity, distance from medical center, employment, socioeconomic status, social support, caregiver status, comorbid conditions, or patient preference. Study Design Cardiac rehabilitation involves exercise training, education, and lifestyle counseling for patients with cardiac conditions. Home and center-based cardiac rehabilitation programs, if followed by patients, are equally effective in their benefits on risk factors, health-related quality of life, death, and clinical events following acute myocardial infarction or coronary revascularization. This study will compare the effectiveness of existing cardiac rehabilitation programs. The San Francisco VA Medical center offers a home-based (but not a center-based) cardiac rehabilitation program to patients who are hospitalized for myocardial infarction, percutaneous coronary intervention, or coronary artery bypass grafting. The other study sites offer a center-based (but not a home-based) cardiac rehabilitation to the same group of patients. All facilities have initiated a comprehensive CR intervention with 7 components:

1. Educate and engage stakeholders (grand rounds, posters, brochures, one-on-one meetings) 2. Automate referrals 3. Standardize templates and stop codes for consults, progress notes, exercise prescription, and the patients' Individualized Treatment Plans 4. Use database to track all patients referred; monitor % enrolled, number of sessions completed. 5. Standardize outcome assessment 6. For CABG patients, provide in-hospital assessment and phase I cardiac rehabilitation; coordinate care with home PT 7. Offer a phase II cardiac rehabilitation program within 6 months of hospitalization. The only difference between the sites is that San Francisco offers a home-based cardiac rehab program and other sites will offer a center-based cardiac rehab program. Research outcomes will include processes of care (% of eligible patients referred, % enrolled, average number of sessions completed, provider knowledge) and patient outcomes (6 minute walk distance, physical activity, quality of life, patient and caregiver satisfaction, re-hospitalization, mortality). The study will evaluate the effectiveness of each CR program on these outcomes. In addition, research participants will be asked to participate in a total of 3 additional study visits. The visits will take place 3 months, 6 months, and 12 months after their enrollment in the cardiac rehabilitation program. Medical records will be reviewed and death certificates will be obtained for up to 10 years of follow-up. Statistical Analysis Plan The outcomes listed will be measured at baseline, and then at 3, 6, and 12 months following first visit to CR. Prior to inferential statistics, we will use descriptive statistics and graphs to explore distributions, univariate outliers, and basic bivariate associations with home- vs. center-based CR. We will then compare baseline characteristics of the two groups to examine differences on key subgroup variables including age and type of procedure resulting in the need for CR. Imbalanced baseline measures will be included as covariates in outcome analyses. To appropriately model these nested data, we will use linear mixed models also called multilevel or hierarchical linear models. Our primary outcomes on which only the patient or their caregiver provide data will be analyzed using a multivariate regression model with a simple random -effects structure with repeated-measures. We anticipate that patients who agree to enroll in home-based CR will be different than those who agree to enroll in center- based CR. Therefore, we will use other variables measured at baseline to construct a propensity score for home- vs. center-based CR, and we will adjust the multivariable analyses for this propensity score. Planned subgroup analyses will include the effects of the following predictors on outcomes: age, gender, race/ethnicity, SES, distance to nearest facility that can provide rehabilitation, and presence of caregivers. We will determine whether these same socio-demographic factors are associated with patient preference. We will also conduct subgroup analyses where we have identical measures from patients and caregivers.

◆